This year I had a few examples where D-790 tests for USCG certified vessels would have generally good test results, then one test would be way, way low. Nobody could figure it out. At IBEX, Joe Parker asked the right questions and figured it out. It’s so obvious that I’m embarrassed. I will be updating my guide to building test samples.
First, most of the testing is done on triaxial roving. I do always have the guys orient and label the 0 degree on the laminate to be tested. The part I was ignoring and Joe got was what happens in the Z direction. (Like Spock said about Khan “he seems to be exhibiting 2 dimensional thinking.”) . The key is to also orient the 0 degree laminate faces away from the sample centroid. I have not before advised on that. On a two layer laminate, the 0 degree goes face down first.  Then the next layer must be face up.  Of course!  If a sample had both 0 degree faces in the centroid, the results would be miserable.